Source code for cosmetic_changes
"""
This module can do slight modifications to tidy a wiki page's source code.
The changes are not supposed to change the look of the rendered wiki page.
If you wish to run this as an stand-alone script, use::
scripts/cosmetic_changes.py
For regular use, it is recommended to put this line into your user config::
cosmetic_changes = True
You may enable cosmetic changes for additional languages by adding the
dictionary cosmetic_changes_enable to your user-config.py. It should contain
a tuple of languages for each site where you wish to enable in addition to
your own langlanguage if cosmetic_changes_mylang_only is True (see below).
Please set your dictionary by adding such lines to your user config::
cosmetic_changes_enable['wikipedia'] = ('de', 'en', 'fr')
There is another config variable: You can set::
cosmetic_changes_mylang_only = False
if you're running a bot on multiple sites and want to do cosmetic changes on
all of them, but be careful if you do.
You may disable cosmetic changes by adding the all unwanted languages to
the `dictionary cosmetic_changes_disable` in your user config file
(`user_config.py`). It should contain a tuple of languages for each site
where you wish to disable cosmetic changes. You may use it with
`cosmetic_changes_mylang_only` is False, but you can also disable your
own language. This also overrides the settings in the dictionary
`cosmetic_changes_enable`. Please set this dictionary by adding such
lines to your user config file::
cosmetic_changes_disable['wikipedia'] = ('de', 'en', 'fr')
You may disable cosmetic changes for a given script by appending the all
unwanted scripts to the list cosmetic_changes_deny_script in your
user-config.py. By default it contains cosmetic_changes.py itself and touch.py.
This overrides all other enabling settings for cosmetic changes. Please modify
the given list by adding such lines to your user-config.py::
cosmetic_changes_deny_script.append('your_script_name_1')
or by adding a list to the given one::
cosmetic_changes_deny_script += ['your_script_name_1',
'your_script_name_2']
"""
#
# (C) Pywikibot team, 2006-2023
#
# Distributed under the terms of the MIT license.
#
from __future__ import annotations
import re
from contextlib import suppress
from enum import IntEnum
from typing import Any
from urllib.parse import urlparse, urlunparse
import pywikibot
from pywikibot import exceptions, textlib
from pywikibot.backports import Callable, Match, Pattern
from pywikibot.site import Namespace
from pywikibot.textlib import (
FILE_LINK_REGEX,
MultiTemplateMatchBuilder,
get_regexes,
)
from pywikibot.tools import first_lower, first_upper
from pywikibot.tools.chars import url2string
try:
import stdnum.isbn as stdnum_isbn
except ImportError:
stdnum_isbn = None
# Subpage templates. Must be in lower case,
# whereas subpage itself must be case sensitive
# This is also used by interwiki.py
# TODO: Maybe move it to family file and implement global instances
moved_links = {
'ar': (['documentation', 'template documentation', 'شرح', 'توثيق'],
'/doc'),
'bn': ('documentation', '/doc'),
'ca': ('ús de la plantilla', '/ús'),
'cs': ('dokumentace', '/doc'),
'da': ('dokumentation', '/doc'),
'de': ('dokumentation', '/Meta'),
'dsb': (['dokumentacija', 'doc'], '/Dokumentacija'),
'en': (['documentation', 'template documentation', 'template doc',
'doc', 'documentation, template'], '/doc'),
'es': (['documentación', 'documentación de plantilla'], '/doc'),
'eu': ('txantiloi dokumentazioa', '/dok'),
'fa': (['documentation', 'template documentation', 'template doc',
'doc', 'توضیحات', 'زیرصفحه توضیحات'], '/doc'),
# fi: no idea how to handle this type of subpage at :Metasivu:
'fi': ('mallineohje', None),
'fr': (['/documentation', 'documentation', 'doc_modèle',
'documentation modèle', 'documentation modèle compliqué',
'documentation modèle en sous-page',
'documentation modèle compliqué en sous-page',
'documentation modèle utilisant les parserfunctions en sous-page',
],
'/Documentation'),
'hsb': (['dokumentacija', 'doc'], '/Dokumentacija'),
'hu': ('sablondokumentáció', '/doc'),
'id': ('template doc', '/doc'),
'ilo': ('documentation', '/doc'),
'ja': ('documentation', '/doc'),
'ka': ('თარგის ინფო', '/ინფო'),
'ko': ('documentation', '/설명문서'),
'ms': ('documentation', '/doc'),
'no': ('dokumentasjon', '/dok'),
'nn': ('dokumentasjon', '/dok'),
'pl': ('dokumentacja', '/opis'),
'pt': (['documentação', '/doc'], '/doc'),
'ro': ('documentaţie', '/doc'),
'ru': ('doc', '/doc'),
'simple': (['documentation',
'template documentation',
'template doc',
'doc',
'documentation, template'], '/doc'),
'sk': ('dokumentácia', '/Dokumentácia'),
'sv': ('dokumentation', '/dok'),
'uk': (['документація', 'doc', 'documentation'], '/Документація'),
'ur': (['دستاویز', 'توثيق', 'شرح', 'توضیحات',
'documentation', 'template doc', 'doc',
'documentation, template'], '/doc'),
'vi': ('documentation', '/doc'),
'zh': (['documentation', 'doc'], '/doc'),
}
# Template which should be replaced or removed.
# Use a list with two entries. The first entry will be replaced by the second.
# Examples:
# For removing {{Foo}}, the list must be:
# ('Foo', None),
#
# The following also works:
# ('Foo', ''),
#
# For replacing {{Foo}} with {{Bar}} the list must be:
# ('Foo', 'Bar'),
#
# This also removes all template parameters of {{Foo}}
# For replacing {{Foo}} with {{Bar}} but keep the template
# parameters in its original order, please use:
# ('Foo', 'Bar\\g<parameters>'),
deprecatedTemplates = {
'wikipedia': {
'de': [
('Belege', 'Belege fehlen\\g<parameters>'),
('Quelle', 'Belege fehlen\\g<parameters>'),
('Quellen', 'Belege fehlen\\g<parameters>'),
('Quellen fehlen', 'Belege fehlen\\g<parameters>'),
],
'ur': [
('Infobox former country',
'خانہ معلومات سابقہ ملک\\g<parameters>'),
('Infobox Former Country',
'خانہ معلومات سابقہ ملک\\g<parameters>'),
],
}
}
[docs]
class CANCEL(IntEnum):
"""Cancel level to ignore exceptions.
If an error occurred and either skips the page or the method
or a single match. ALL raises the exception.
.. versionadded:: 6.3
"""
ALL = 0
PAGE = 1
METHOD = 2
MATCH = 3
def _format_isbn_match(match: Match[str], strict: bool = True) -> str:
"""Helper function to validate and format a single matched ISBN."""
if not stdnum_isbn:
raise NotImplementedError(
'ISBN functionality not available. Install stdnum package.')
isbn = match['code']
try:
stdnum_isbn.validate(isbn)
except stdnum_isbn.ValidationError as e:
if strict:
raise
pywikibot.log(f'ISBN "{isbn}" validation error: {e}')
return isbn
return stdnum_isbn.format(isbn)
def _reformat_ISBNs(text: str, strict: bool = True) -> str:
"""Helper function to normalise ISBNs in text.
:raises Exception: Invalid ISBN encountered when strict enabled
"""
return textlib.reformat_ISBNs(
text, lambda match: _format_isbn_match(match, strict=strict))
[docs]
class CosmeticChangesToolkit:
"""Cosmetic changes toolkit.
.. versionchanged:: 7.0
`from_page()` method was removed
"""
def __init__(self, page: pywikibot.page.BasePage, *,
show_diff: bool = False,
ignore: IntEnum = CANCEL.ALL) -> None:
"""Initializer.
.. versionchanged:: 5.2
instantiate the CosmeticChangesToolkit from a page object;
only allow keyword arguments except for page parameter;
`namespace` and `pageTitle` parameters are deprecated
.. versionchanged:: 7.0
`namespace` and `pageTitle` parameters were removed
:param page: the Page object containing the text to be modified
:param show_diff: show difference after replacements
:param ignore: ignores if an error occurred and either skips the page
or only that method. It can be set one of the CANCEL constants
"""
self.site = page.site
self.title = page.title()
self.namespace = page.namespace()
self.show_diff = show_diff
self.template = (self.namespace == Namespace.TEMPLATE)
self.talkpage = self.namespace >= 0 and self.namespace % 2 == 1
self.ignore = ignore
self.common_methods = [
self.commonsfiledesc,
self.fixSelfInterwiki,
self.standardizePageFooter,
self.fixSyntaxSave,
self.cleanUpLinks,
self.cleanUpSectionHeaders,
self.putSpacesInLists,
self.translateAndCapitalizeNamespaces,
self.translateMagicWords,
self.replaceDeprecatedTemplates,
self.resolveHtmlEntities,
self.removeEmptySections,
self.removeUselessSpaces,
self.removeNonBreakingSpaceBeforePercent,
self.fixHtml,
self.fixReferences,
self.fixStyle,
self.fixTypo,
self.fixArabicLetters,
]
if stdnum_isbn:
self.common_methods.append(self.fix_ISBN)
[docs]
def safe_execute(self, method: Callable[[str], str], text: str) -> str:
"""Execute the method and catch exceptions if enabled."""
result = None
try:
result = method(text)
except Exception as e:
if self.ignore == CANCEL.METHOD:
pywikibot.warning('Unable to perform "{}" on "{}"!'
.format(method.__name__, self.title))
pywikibot.error(e)
else:
raise
return text if result is None else result
def _change(self, text: str) -> str:
"""Execute all clean up methods."""
for method in self.common_methods:
text = self.safe_execute(method, text)
return text
[docs]
def change(self, text: str) -> bool | str:
"""Execute all clean up methods and catch errors if activated."""
try:
new_text = self._change(text)
except Exception as e:
if self.ignore == CANCEL.PAGE:
pywikibot.warning('Skipped "{}", because an error occurred.'
.format(self.title))
pywikibot.error(e)
return False
raise
else:
if self.show_diff:
pywikibot.showDiff(text, new_text)
return new_text
[docs]
def fixSelfInterwiki(self, text: str) -> str:
"""
Interwiki links to the site itself are displayed like local links.
Remove their language code prefix.
"""
if not self.talkpage and pywikibot.calledModuleName() != 'interwiki':
interwikiR = re.compile(r'\[\[(?: *:)? *{} *: *([^\[\]\n]*)\]\]'
.format(self.site.code))
text = interwikiR.sub(r'[[\1]]', text)
return text
[docs]
def translateAndCapitalizeNamespaces(self, text: str) -> str:
"""Use localized namespace names.
.. versionchanged:: 7.4
No longer expect a specific namespace alias for File:
"""
# arz uses English stylish codes
if self.site.sitename == 'wikipedia:arz':
return text
# wiki links aren't parsed here.
exceptions = ['nowiki', 'comment', 'math', 'pre']
for namespace in self.site.namespaces.values():
if namespace == Namespace.MAIN:
# skip main (article) namespace
continue
# a clone is needed. Won't change the namespace dict
namespaces = list(namespace)
if (namespace == Namespace.FILE
and self.site.family.name == 'wikipedia'):
if self.site.code in ('en', 'fr'):
# do not change "Image" on en-wiki and fr-wiki
with suppress(ValueError):
namespaces.remove('Image')
if self.site.code == 'hu':
# do not change "Kép" on hu-wiki
with suppress(ValueError):
namespaces.remove('Kép')
elif self.site.code == 'pt':
# use "Imagem" by default on pt-wiki (per T57242)
with suppress(ValueError):
namespaces.insert(
0, namespaces.pop(namespaces.index('Imagem')))
# final namespace variant
final_ns = namespaces.pop(0)
if namespace in (Namespace.USER, Namespace.USER_TALK):
# skip localized user namespace, maybe gender is used
namespaces = ['User'
if namespace == Namespace.USER
else 'User talk']
# lowerspaced and underscored namespaces
for i, item in enumerate(namespaces):
item = item.replace(' ', '[ _]')
item = f'[{item[0]}{item[0].lower()}]' + item[1:]
namespaces[i] = item
namespaces.append(first_lower(final_ns))
if final_ns and namespaces:
if (self.site.sitename == 'wikipedia:pt'
and namespace == Namespace.FILE):
# only change on these file extensions (per T57242)
extensions = ('png', 'gif', 'jpg', 'jpeg', 'svg', 'tiff',
'tif')
text = textlib.replaceExcept(
text,
r'\[\[\s*({}) *:(?P<name>[^\|\]]*?\.({}))'
r'(?P<label>.*?)\]\]'
.format('|'.join(namespaces), '|'.join(extensions)),
fr'[[{final_ns}:\g<name>\g<label>]]',
exceptions)
else:
text = textlib.replaceExcept(
text,
r'\[\[\s*({}) *:(?P<nameAndLabel>.*?)\]\]'
.format('|'.join(namespaces)),
fr'[[{final_ns}:\g<nameAndLabel>]]',
exceptions)
return text
[docs]
def translateMagicWords(self, text: str) -> str:
"""Use localized magic words."""
# not wanted at ru
# arz uses English stylish codes
# no need to run on English wikis
if self.site.code in ['arz', 'en', 'ru']:
return text
def init_cache() -> None:
for magicword in ('img_thumbnail', 'img_left', 'img_center',
'img_right', 'img_none', 'img_framed',
'img_frameless', 'img_border', 'img_upright',
'img_baseline', 'img_sub', 'img_super',
'img_top', 'img_text_top', 'img_middle',
'img_bottom', 'img_text_bottom'):
aliases = self.site.getmagicwords(magicword)
if len(aliases) > 1:
cache.update((alias, aliases[0]) for alias in aliases[1:]
if '$1' not in alias)
if not cache:
cache[False] = True # signal there is nothing to replace
def replace_magicword(match: Match[str]) -> str:
if cache.get(False):
return match.group()
split = match.group().split('|')
if len(split) == 1:
return match.group()
if not cache:
init_cache()
# push ']]' out and re-add below
split[-1] = split[-1][:-2]
return '{}|{}]]'.format(
split[0], '|'.join(cache.get(x.strip(), x) for x in split[1:]))
cache: dict[bool | str, Any] = {}
exceptions = ['comment', 'nowiki', 'pre', 'syntaxhighlight']
regex = re.compile(
FILE_LINK_REGEX % '|'.join(self.site.namespaces[6]),
flags=re.X)
return textlib.replaceExcept(
text, regex, replace_magicword, exceptions)
[docs]
def cleanUpLinks(self, text: str) -> str:
"""Tidy up wikilinks found in a string.
This function will:
* Replace underscores with spaces
* Move leading and trailing spaces out of the wikilink and into the
surrounding text
* Convert URL-encoded characters into Unicode-encoded characters
* Move trailing characters out of the link and make the link without
using a pipe, if possible
* Capitalize the article title of the link, if appropriate
.. versionchanged:: 8.4
Convert URL-encoded characters if a link is an interwiki link
or different from main namespace.
:param text: string to perform the clean-up on
:return: text with tidied wikilinks
"""
# helper function which works on one link and either returns it
# unmodified, or returns a replacement.
def handleOneLink(match: Match[str]) -> str:
# Convert URL-encoded characters to str
titleWithSection = url2string(match['titleWithSection'],
encodings=self.site.encodings())
label = match['label']
trailingChars = match['linktrail']
newline = match['newline']
# entire link but convert URL-encoded text
oldlink = url2string(match.group(),
encodings=self.site.encodings())
is_interwiki = None
with suppress(exceptions.ServerError):
is_interwiki = self.site.isInterwikiLink(titleWithSection)
if is_interwiki is not False:
return oldlink
# The link looks like this:
# [[page_title|link_text]]trailing_chars
# We only work on namespace 0 because pipes and linktrails work
# differently for images and categories.
page = pywikibot.Page(pywikibot.Link(titleWithSection, self.site))
in_main_namespace = None
with suppress(exceptions.InvalidTitleError):
in_main_namespace = page.namespace() == Namespace.MAIN
if not in_main_namespace:
return oldlink
# Replace underlines by spaces, also multiple underlines
titleWithSection = re.sub('_+', ' ', titleWithSection)
# Remove double spaces
titleWithSection = re.sub(' +', ' ', titleWithSection)
# Remove unnecessary leading spaces from title,
# but remember if we did this because we eventually want
# to re-add it outside of the link later.
titleLength = len(titleWithSection)
titleWithSection = titleWithSection.lstrip()
hadLeadingSpaces = len(titleWithSection) != titleLength
hadTrailingSpaces = False
# Remove unnecessary trailing spaces from title,
# but remember if we did this because it may affect
# the linktrail and because we eventually want to
# re-add it outside of the link later.
if not trailingChars:
titleLength = len(titleWithSection)
titleWithSection = titleWithSection.rstrip()
hadTrailingSpaces = len(titleWithSection) != titleLength
if not titleWithSection:
# just skip empty links.
return match.group()
# Remove unnecessary initial and final spaces from label.
# Please note that some editors prefer spaces around pipes.
# (See [[en:Wikipedia:Semi-bots]]). We remove them anyway.
if label is not None:
# Remove unnecessary leading spaces from label,
# but remember if we did this because we want
# to re-add it outside of the link later.
labelLength = len(label)
label = label.lstrip()
hadLeadingSpaces = len(label) != labelLength
# Remove unnecessary trailing spaces from label,
# but remember if we did this because it affects
# the linktrail.
if not trailingChars:
labelLength = len(label)
label = label.rstrip()
hadTrailingSpaces = len(label) != labelLength
else:
label = titleWithSection
if trailingChars:
label += trailingChars
if self.site.siteinfo['case'] == 'first-letter':
firstcase_title = first_lower(titleWithSection)
firstcase_label = first_lower(label)
else:
firstcase_title = titleWithSection
firstcase_label = label
if firstcase_label == firstcase_title:
newLink = f'[[{label}]]'
# Check if we can create a link with trailing characters
# instead of a pipelink
elif (firstcase_label.startswith(firstcase_title)
and trailR.sub('', label[len(titleWithSection):]) == ''):
newLink = '[[{}]]{}'.format(label[:len(titleWithSection)],
label[len(titleWithSection):])
else:
# Try to capitalize the first letter of the title.
# Not useful for languages that don't capitalize nouns.
# TODO: Add a configuration variable for each site,
# which determines if the link target is written in
# uppercase
if self.site.sitename == 'wikipedia:de':
titleWithSection = first_upper(titleWithSection)
newLink = f'[[{titleWithSection}|{label}]]'
# re-add spaces that were pulled out of the link.
# Examples:
# text[[ title ]]text -> text [[title]] text
# text[[ title | name ]]text -> text [[title|name]] text
# text[[ title |name]]text -> text[[title|name]]text
# text[[title| name]]text -> text [[title|name]]text
if hadLeadingSpaces and not newline:
newLink = ' ' + newLink
if hadTrailingSpaces:
newLink += ' '
if newline:
newLink = newline + newLink
return newLink
trailR = re.compile(self.site.linktrail())
# The regular expression which finds links. Results consist of four groups:
# group <newline> depends whether the links starts with a new line.
# group <titleWithSection> is the page title and section, that is,
# everything before | or ]. It'll include the # to make life easier for us.
# group <label> is the alternative link title between | and ].
# group <linktrail> is the link trail after ]] which are part of the word.
# note that the definition of 'letter' varies from language to language.
linkR = re.compile(
r'(?P<newline>[\n]*)\[\[(?P<titleWithSection>[^\]\|]+)'
r'(\|(?P<label>[^\]\|]*))?\]\](?P<linktrail>'
+ self.site.linktrail() + ')')
text = textlib.replaceExcept(text, linkR, handleOneLink,
['comment', 'math', 'nowiki', 'pre',
'startspace'])
return text
[docs]
def resolveHtmlEntities(self, text: str) -> str:
"""Replace HTML entities with string."""
ignore = [
38, # Ampersand (&)
39, # Single quotation mark (") per T26093
60, # Less than (<)
62, # Greater than (>)
91, # Opening square bracket ([)
# - sometimes used intentionally inside links
93, # Closing square bracket (])
# - used intentionally inside links
124, # Vertical bar (|)
# - used intentionally in navigation bar templates on w:de
160, # Non-breaking space ( )
# - not supported by Firefox textareas
173, # Soft-hypen (­) - enable editing
8206, # Left-to-right mark (<r;)
8207, # Right-to-left mark (&rtl;)
]
if self.template:
ignore.append(32) # Space ( )
ignore.append(58) # Colon (:)
# TODO: T254350 - what other extension tags should be avoided?
# (graph, math, score, timeline, etc.)
text = pywikibot.html2unicode(
text, ignore=ignore, exceptions=['comment', 'syntaxhighlight'])
return text
[docs]
def removeEmptySections(self, text: str) -> str:
"""Cleanup empty sections."""
# userspace contains article stubs without nobots/in use templates
if self.namespace == Namespace.USER:
return text
skippings = ['comment', 'category']
skip_regexes = get_regexes(skippings, self.site)
# site defined templates
skip_templates = {
'cs': ('Pahýl[ _]část',), # stub section
}
if self.site.code in skip_templates:
for template in skip_templates[self.site.code]:
skip_regexes.append(
re.compile(r'\{\{\s*%s\s*\}\}' % template, re.I))
# empty lists
skip_regexes.append(re.compile(r'(?m)^[\*#] *$'))
# get stripped sections
stripped_text = textlib.removeLanguageLinks(text, self.site, '\n')
for reg in skip_regexes:
stripped_text = reg.sub(r'', stripped_text)
strip_sections = textlib.extract_sections(
stripped_text, self.site).sections
# get proper sections
header, sections, footer = textlib.extract_sections(text, self.site)
# iterate stripped sections and create a new page body
new_body: list[textlib.Section] = []
for i, strip_section in enumerate(strip_sections):
current_dep = sections[i].level
try:
next_dep = sections[i + 1].level
except IndexError:
next_dep = 0
if strip_section.content.strip() or current_dep < next_dep:
new_body.extend(sections[i])
return header + ''.join(new_body) + footer
[docs]
def removeUselessSpaces(self, text: str) -> str:
"""Cleanup multiple or trailing spaces."""
exceptions = ['comment', 'math', 'nowiki', 'pre', 'syntaxhighlight',
'startspace', 'table']
if self.site.sitename != 'wikipedia:cs':
exceptions.append('template')
text = textlib.replaceExcept(text, r'(?m)[\t ]+( |$)', r'\1',
exceptions, site=self.site)
return text
[docs]
def removeNonBreakingSpaceBeforePercent(self, text: str) -> str:
"""
Remove a non-breaking space between number and percent sign.
Newer MediaWiki versions automatically place a non-breaking space in
front of a percent sign, so it is no longer required to place it
manually.
"""
text = textlib.replaceExcept(
text, r'(\d)&(?:nbsp|#160|#x[Aa]0);%', r'\1 %', ['timeline'])
return text
[docs]
def cleanUpSectionHeaders(self, text: str) -> str:
"""
Add a space between the equal signs and the section title.
Example::
==Section title==
becomes::
== Section title ==
.. note:: This space is recommended in the syntax help on the
English and German Wikipedias. It is not wanted on Lojban and
English Wiktionaries (:phab:`T168399`, :phab:`T169064`) and
it might be that it is not wanted on other wikis. If there
are any complaints, please file a bug report.
"""
if self.site.sitename in ['wiktionary:jbo', 'wiktionary:en']:
return text
return textlib.replaceExcept(
text,
r'(?m)^(={1,6})[ \t]*(?P<title>.*[^\s=])[ \t]*\1[ \t]*\r?\n',
r'\1 \g<title> \1\n',
['comment', 'math', 'nowiki', 'pre'])
[docs]
def putSpacesInLists(self, text: str) -> str:
"""
Add a space between the * or # and the text.
.. note:: This space is recommended in the syntax help on the
English, German and French Wikipedias. It might be that it
is not wanted on other wikis. If there are any complaints,
please file a bug report.
"""
if not self.template:
exceptions = ['comment', 'math', 'nowiki', 'pre',
'syntaxhighlight', 'template', 'timeline',
self.site.redirect_regex]
text = textlib.replaceExcept(
text,
r'(?m)'
r'^(?P<bullet>[:;]*(\*+|#+)[:;\*#]*)(?P<char>[^\s\*#:;].+?)',
r'\g<bullet> \g<char>',
exceptions)
return text
[docs]
def replaceDeprecatedTemplates(self, text: str) -> str:
"""Replace deprecated templates."""
exceptions = ['comment', 'math', 'nowiki', 'pre']
builder = MultiTemplateMatchBuilder(self.site)
if self.site.family.name in deprecatedTemplates \
and self.site.code in deprecatedTemplates[self.site.family.name]:
for template in deprecatedTemplates[
self.site.family.name][self.site.code]:
old, new = template
if new is None:
new = ''
else:
new = '{{%s}}' % new
text = textlib.replaceExcept(
text,
builder.pattern(old),
new, exceptions)
return text
# from fixes.py
[docs]
def fixSyntaxSave(self, text: str) -> str:
"""Convert weblinks to wikilink, fix link syntax."""
def replace_link(match: Match[str]) -> str:
"""Create a string to replace a single link."""
replacement = '[['
if re.match(
r'(?:{}):'.format('|'.join(
(*self.site.namespaces[Namespace.FILE],
*self.site.namespaces[Namespace.CATEGORY]))),
match['link']):
replacement += ':'
replacement += match['link']
if match['title']:
replacement += '|' + match['title']
return replacement + ']]'
exceptions = ['comment', 'math', 'nowiki', 'pre', 'startspace',
'syntaxhighlight']
# link to the wiki working on
# Only use suffixes for article paths
for suffix in self.site._interwiki_urls(True):
http_url = self.site.base_url(suffix, 'http')
if self.site.protocol() == 'http':
https_url = None
else:
https_url = self.site.base_url(suffix, 'https')
# compare strings without the protocol, if they are empty support
# also no prefix (//en.wikipedia.org/…)
http = urlparse(http_url)
https = urlparse(https_url)
if https_url is not None and http.netloc == https.netloc:
urls = ['(?:https?:)?'
+ re.escape(urlunparse(('', *http[1:])))]
else:
urls = [re.escape(url) for url in (http_url, https_url)
if url is not None]
for url in urls:
# unescape {} placeholder
url = url.replace(r'\{\}', '{title}')
# Only include links which don't include the separator
# as the wikilink won't support additional parameters
separator = '?&' if '?' in suffix else '?'
# Match first a non space in the title to prevent that multiple
# spaces at the end without title will be matched by it
title_regex = (r'(?P<link>[^{sep}]+?)'
r'(\s+(?P<title>[^\s].*?))'
.format(sep=separator))
url_regex = fr'\[\[?{url}?\s*\]\]?'
text = textlib.replaceExcept(
text,
url_regex.format(title=title_regex),
replace_link, exceptions, site=self.site)
# external link in/starting with double brackets
text = textlib.replaceExcept(
text,
r'\[\[(?P<url>https?://[^\]]+?)\]\]?',
r'[\g<url>]', exceptions, site=self.site)
# external link and description separated by a pipe, with
# whitespace in front of the pipe, so that it is clear that
# the dash is not a legitimate part of the URL.
text = textlib.replaceExcept(
text,
r'\[(?P<url>https?://[^\|\] \r\n]+?) +\| *(?P<label>[^\|\]]+?)\]',
r'[\g<url> \g<label>]', exceptions)
# dash in external link, where the correct end of the URL can
# be detected from the file extension. It is very unlikely that
# this will cause mistakes.
extensions = [fr'\.{ext}'
for ext in ['pdf', 'html?', 'php', 'aspx?', 'jsp']]
text = textlib.replaceExcept(
text,
r'\[(?P<url>https?://[^\|\] ]+?(' + '|'.join(extensions) + r')) *'
r'\| *(?P<label>[^\|\]]+?)\]',
r'[\g<url> \g<label>]', exceptions)
return text
[docs]
def fixHtml(self, text: str) -> str:
"""Relace html markups with wikitext markups."""
def replace_header(match: Match[str]) -> str:
"""Create a header string for replacing."""
depth = int(match[1])
return r'{0} {1} {0}'.format('=' * depth, match[2])
# Everything case-insensitive (?i)
# Keep in mind that MediaWiki automatically converts <br> to <br />
exceptions = ['comment', 'math', 'nowiki', 'pre', 'startspace',
'syntaxhighlight']
text = textlib.replaceExcept(text, r'(?i)<(b|strong)>(.*?)</\1>',
r"'''\2'''", exceptions, site=self.site)
text = textlib.replaceExcept(text, r'(?i)<(i|em)>(.*?)</\1>',
r"''\2''", exceptions, site=self.site)
# horizontal line without attributes in a single line
text = textlib.replaceExcept(text, r'(?i)([\r\n])<hr[ /]*>([\r\n])',
r'\1----\2', exceptions)
# horizontal line with attributes; can't be done with wiki syntax
# so we only make it XHTML compliant
text = textlib.replaceExcept(text, r'(?i)<hr ([^>/]+?)>',
r'<hr \1 />',
exceptions)
# a header where only spaces are in the same line
text = textlib.replaceExcept(
text,
r'(?i)(?<=[\r\n]) *<h([1-7])> *([^<]+?) *</h\1> *(?=[\r\n])',
replace_header,
exceptions)
# TODO: maybe we can make the bot replace <p> tags with \r\n's.
return text
[docs]
def fixReferences(self, text: str) -> str:
"""Fix references tags."""
# See also
# https://en.wikipedia.org/wiki/User:AnomieBOT/source/tasks/OrphanReferenceFixer.pm
exceptions = ['comment', 'math', 'nowiki', 'pre', 'syntaxhighlight',
'startspace']
# it should be name = " or name=" NOT name ="
text = re.sub(r'(?i)<ref +name(= *| *=)"', r'<ref name="', text)
# remove empty <ref/>-tag
text = textlib.replaceExcept(text,
r'(?i)(<ref\s*/>|<ref *>\s*</ref>)',
r'', exceptions)
text = textlib.replaceExcept(text,
r'(?i)<ref\s+([^>]+?)\s*>\s*</ref>',
r'<ref \1/>', exceptions)
return text
[docs]
def fixStyle(self, text: str) -> str:
"""Convert prettytable to wikitable class."""
exceptions = ['comment', 'math', 'nowiki', 'pre', 'startspace',
'syntaxhighlight']
if self.site.code in ('de', 'en'):
text = textlib.replaceExcept(text,
r'(class="[^"]*)prettytable([^"]*")',
r'\1wikitable\2', exceptions)
return text
[docs]
def fixTypo(self, text: str) -> str:
"""Fix units."""
exceptions: list[str | Pattern[str]] = [
'comment',
'gallery',
'hyperlink',
'interwiki',
'link',
'nowiki',
'math',
'pre',
'startspace',
'syntaxhighlight',
]
# change <number> ccm -> <number> cm³
text = textlib.replaceExcept(text, r'(\d)\s*(?: )?ccm',
r'\1 cm³', exceptions,
site=self.site)
# Solve wrong Nº sign with °C or °F
# additional exception requested on fr-wiki for this stuff
pattern = re.compile('«.*?»')
exceptions.append(pattern)
text = textlib.replaceExcept(text, r'(\d)\s*(?: )?[º°]([CF])',
r'\1 °\2', exceptions,
site=self.site)
text = textlib.replaceExcept(text, 'º([CF])', '°' + r'\1',
exceptions,
site=self.site)
return text
[docs]
def fixArabicLetters(self, text: str) -> str:
"""Fix Arabic and Persian letters."""
if self.site.code not in ['ckb', 'fa']:
return text
exceptions: list[str | Pattern[str]] = [
'file',
'gallery',
'hyperlink',
'interwiki',
'inputbox',
# FIXME: but changes letters inside wikilinks
# 'link',
'math',
'pre',
'template',
'timeline',
'ref',
'startspace',
'syntaxhighlight',
]
digits = textlib.NON_LATIN_DIGITS
faChrs = 'ءاآأإئؤبپتثجچحخدذرزژسشصضطظعغفقکگلمنوهیةيك' + digits['fa']
# not to let bot edits in latin content
exceptions.append(re.compile('[^{fa}] *?"*? *?, *?[^{fa}]'
.format(fa=faChrs)))
text = textlib.replaceExcept(text, ',', '،', exceptions,
site=self.site)
if self.site.code == 'ckb':
text = textlib.replaceExcept(text,
'\u0647([.\u060c_<\\]\\s])',
'\u06d5\\1', exceptions,
site=self.site)
text = textlib.replaceExcept(text, 'ه\u200c', 'ە', exceptions,
site=self.site)
text = textlib.replaceExcept(text, 'ه', 'ھ', exceptions,
site=self.site)
text = textlib.replaceExcept(text, 'ك', 'ک', exceptions,
site=self.site)
text = textlib.replaceExcept(text, '[ىي]', 'ی', exceptions,
site=self.site)
return text
[docs]
def commonsfiledesc(self, text: str) -> str:
"""
Clean up file descriptions on Wikimedia Commons.
It works according to [1] and works only on pages in the file
namespace on Wikimedia Commons.
[1]:
https://commons.wikimedia.org/wiki/Commons:Tools/pywiki_file_description_cleanup
"""
if (self.site.sitename != 'commons:commons'
or self.namespace == Namespace.FILE):
return text
# section headers to {{int:}} versions
exceptions = ['comment', 'includeonly', 'math', 'noinclude', 'nowiki',
'pre', 'syntaxhighlight', 'ref', 'timeline']
text = textlib.replaceExcept(text,
r'([\r\n]|^)\=\= *Summary *\=\=',
r'\1== {{int:filedesc}} ==',
exceptions, True)
text = textlib.replaceExcept(
text,
r'([\r\n])\=\= *\[\[Commons:Copyright tags\|Licensing\]\]: *\=\=',
r'\1== {{int:license-header}} ==', exceptions, True)
text = textlib.replaceExcept(
text,
r'([\r\n])'
r'\=\= *(Licensing|License information|{{int:license}}) *\=\=',
r'\1== {{int:license-header}} ==', exceptions, True)
# frequent field values to {{int:}} versions
text = textlib.replaceExcept(
text,
r'([\r\n]\|[Ss]ource *\= *)'
r'(?:[Oo]wn work by uploader|[Oo]wn work|[Ee]igene [Aa]rbeit) *'
r'([\r\n])',
r'\1{{own}}\2', exceptions, True)
text = textlib.replaceExcept(
text,
r'(\| *Permission *\=) *(?:[Ss]ee below|[Ss]iehe unten) *([\r\n])',
r'\1\2', exceptions, True)
# added to transwikied pages
text = textlib.replaceExcept(text, r'__NOTOC__', '', exceptions, True)
# tracker element for js upload form
text = textlib.replaceExcept(
text,
r'<!-- *{{ImageUpload\|(?:full|basic)}} *-->',
'', exceptions[1:], True)
text = textlib.replaceExcept(text, r'{{ImageUpload\|(?:basic|full)}}',
'', exceptions, True)
# duplicated section headers
text = textlib.replaceExcept(
text,
r'([\r\n]|^)\=\= *{{int:filedesc}} *\=\=(?:[\r\n ]*)\=\= *'
r'{{int:filedesc}} *\=\=',
r'\1== {{int:filedesc}} ==', exceptions, True)
text = textlib.replaceExcept(
text,
r'([\r\n]|^)\=\= *{{int:license-header}} *\=\=(?:[\r\n ]*)'
r'\=\= *{{int:license-header}} *\=\=',
r'\1== {{int:license-header}} ==', exceptions, True)
return text
[docs]
def fix_ISBN(self, text: str) -> str:
"""Hyphenate ISBN numbers."""
return _reformat_ISBNs(text, strict=self.ignore != CANCEL.MATCH)